User Details
User Details
- User Since
- Dec 29 2019, 7:44 PM (125 w, 5 d)
Oct 29 2021
Oct 29 2021
asaadaldien committed rG813fa79c151c: Don't drop in_bounds when vector-transfer-collapse-inner-most-dims (authored by asaadaldien).
Don't drop in_bounds when vector-transfer-collapse-inner-most-dims
Oct 29 2021, 9:08 AM · Restricted Project
asaadaldien updated the diff for D112772: Don't drop in_bounds when vector-transfer-collapse-inner-most-dims.
comments
Oct 29 2021, 8:33 AM · Restricted Project
Oct 28 2021
Oct 28 2021
asaadaldien added a reviewer for D112772: Don't drop in_bounds when vector-transfer-collapse-inner-most-dims: ThomasRaoux.
Oct 28 2021, 5:41 PM · Restricted Project
asaadaldien requested review of D112772: Don't drop in_bounds when vector-transfer-collapse-inner-most-dims.
Oct 28 2021, 5:40 PM · Restricted Project
Oct 26 2021
Oct 26 2021
Oct 26 2021, 1:02 PM · Restricted Project
asaadaldien updated the summary of D112570: Use vectorShape instead of vectorWidth.
Oct 26 2021, 12:22 PM · Restricted Project
asaadaldien added a comment to D112566: [mlir] Allow polynomial approximations for N-d vectors..
This diff broke the build, should be fixed in https://reviews.llvm.org/D112570
Oct 26 2021, 12:20 PM · Restricted Project
asaadaldien requested review of D112570: Use vectorShape instead of vectorWidth.
Oct 26 2021, 12:18 PM · Restricted Project
Oct 21 2021
Oct 21 2021
asaadaldien committed rG21f9e4a1ed91: Avoid infinity arithmetics when computing exp approximations (authored by asaadaldien).
Avoid infinity arithmetics when computing exp approximations
Oct 21 2021, 10:09 AM · Restricted Project
asaadaldien updated the diff for D112115: Avoid infinity arithmetics when computing exp approximations.
check vecor type
Oct 21 2021, 8:06 AM · Restricted Project
asaadaldien updated the diff for D112115: Avoid infinity arithmetics when computing exp approximations.
remove XFAIL
Oct 21 2021, 7:58 AM · Restricted Project
Oct 20 2021
Oct 20 2021
asaadaldien committed rGa3dd4e777095: Drop transfer_read inner most unit dimensions (authored by asaadaldien).
Drop transfer_read inner most unit dimensions
Oct 20 2021, 12:27 PM · Restricted Project
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
rebase...
Oct 20 2021, 10:33 AM · Restricted Project
asaadaldien updated the summary of D112115: Avoid infinity arithmetics when computing exp approximations.
Oct 20 2021, 9:29 AM · Restricted Project
asaadaldien added a reviewer for D112115: Avoid infinity arithmetics when computing exp approximations: ezhulenev.
Oct 20 2021, 9:28 AM · Restricted Project
Oct 19 2021
Oct 19 2021
asaadaldien requested review of D112115: Avoid infinity arithmetics when computing exp approximations.
Oct 19 2021, 6:07 PM · Restricted Project
Oct 18 2021
Oct 18 2021
asaadaldien committed rGb0c4aaff24fb: Allow only valid vector.shape_cast transitive folding (authored by asaadaldien).
Allow only valid vector.shape_cast transitive folding
Oct 18 2021, 7:58 AM · Restricted Project
Oct 13 2021
Oct 13 2021
asaadaldien added inline comments to D111722: [mlir][linalg] Add a pattern to vectorize convolution ops.
Oct 13 2021, 10:42 AM · Restricted Project
asaadaldien added inline comments to D111722: [mlir][linalg] Add a pattern to vectorize convolution ops.
Oct 13 2021, 10:36 AM · Restricted Project
asaadaldien updated the diff for D111473: Allow only valid vector.shape_cast transitive folding.
Rebase...
Oct 13 2021, 9:49 AM · Restricted Project
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
update test pass name
Oct 13 2021, 9:31 AM · Restricted Project
Oct 11 2021
Oct 11 2021
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
Drop leading dim in test case
Oct 11 2021, 3:20 PM · Restricted Project
asaadaldien added inline comments to D111473: Allow only valid vector.shape_cast transitive folding.
Oct 11 2021, 2:25 PM · Restricted Project
asaadaldien updated the diff for D111473: Allow only valid vector.shape_cast transitive folding.
don't fold A -> B -> C when rank(A) == rank(C) && type(A) != type(C)
Oct 11 2021, 2:24 PM · Restricted Project
asaadaldien updated the diff for D111473: Allow only valid vector.shape_cast transitive folding.
Assert src rank != result rank
Oct 11 2021, 2:09 PM · Restricted Project
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
Use isMinorIdentitiy
Oct 11 2021, 1:30 PM · Restricted Project
asaadaldien added inline comments to D111561: Drop transfer_read inner most unit dimensions.
Oct 11 2021, 1:08 PM · Restricted Project
asaadaldien added inline comments to D111561: Drop transfer_read inner most unit dimensions.
Oct 11 2021, 1:02 PM · Restricted Project
asaadaldien added inline comments to D111473: Allow only valid vector.shape_cast transitive folding.
Oct 11 2021, 12:44 PM · Restricted Project
asaadaldien added inline comments to D111561: Drop transfer_read inner most unit dimensions.
Oct 11 2021, 12:40 PM · Restricted Project
asaadaldien added inline comments to D111561: Drop transfer_read inner most unit dimensions.
Oct 11 2021, 12:38 PM · Restricted Project
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
Comments..
Oct 11 2021, 12:38 PM · Restricted Project
asaadaldien added a reviewer for D111561: Drop transfer_read inner most unit dimensions: ThomasRaoux.
Oct 11 2021, 10:32 AM · Restricted Project
asaadaldien updated the diff for D111561: Drop transfer_read inner most unit dimensions.
add comment
Oct 11 2021, 10:31 AM · Restricted Project
asaadaldien requested review of D111561: Drop transfer_read inner most unit dimensions.
Oct 11 2021, 10:24 AM · Restricted Project
Oct 9 2021
Oct 9 2021
asaadaldien retitled D111473: Allow only valid vector.shape_cast transitive folding from Allow only valid transitive folding to Allow only valid vector.shape_cast transitive folding.
Oct 9 2021, 4:02 PM · Restricted Project
Oct 8 2021
Oct 8 2021
asaadaldien requested review of D111473: Allow only valid vector.shape_cast transitive folding.
Oct 8 2021, 4:45 PM · Restricted Project
Sep 13 2021
Sep 13 2021
asaadaldien accepted D109692: Reorder mmt4d shapes:.
Sep 13 2021, 11:49 AM · Restricted Project
Aug 3 2021
Aug 3 2021
Reorder mmt4d r.h.s operand layout
Aug 3 2021, 10:22 AM · Restricted Project
Aug 2 2021
Aug 2 2021
asaadaldien updated the diff for D107307: Reorder mmt4d r.h.s operand layout.
update TensorDef
Aug 2 2021, 1:36 PM · Restricted Project
asaadaldien requested review of D107307: Reorder mmt4d r.h.s operand layout.
Aug 2 2021, 1:34 PM · Restricted Project
We are adding this pattern to IREE for now..
Aug 2 2021, 1:29 PM · Restricted Project
Jul 29 2021
Jul 29 2021
Rorder mmt4d iteration domain
Jul 29 2021, 11:32 AM · Restricted Project
asaadaldien updated the diff for D107003: Rorder mmt4d iteration domain.
Rebase and fix commit msg
Jul 29 2021, 9:58 AM · Restricted Project
Jul 28 2021
Jul 28 2021
Jul 28 2021, 1:57 PM · Restricted Project
asaadaldien requested review of D107003: Rorder mmt4d iteration domain.
Jul 28 2021, 1:56 PM · Restricted Project
Jul 14 2021
Jul 14 2021
asaadaldien added inline comments to D106006: Add a pattern to convert static linalg.matmul -> linalg.mmt4d.
Jul 14 2021, 2:00 PM · Restricted Project
asaadaldien added inline comments to D106006: Add a pattern to convert static linalg.matmul -> linalg.mmt4d.
Jul 14 2021, 1:41 PM · Restricted Project
asaadaldien updated the summary of D106006: Add a pattern to convert static linalg.matmul -> linalg.mmt4d.
Jul 14 2021, 12:35 PM · Restricted Project
asaadaldien added a reviewer for D106006: Add a pattern to convert static linalg.matmul -> linalg.mmt4d: Benoit.
Jul 14 2021, 12:34 PM · Restricted Project
asaadaldien requested review of D106006: Add a pattern to convert static linalg.matmul -> linalg.mmt4d.
Jul 14 2021, 12:33 PM · Restricted Project
Jul 1 2021
Jul 1 2021
Add linalg.mmt4d named op
Jul 1 2021, 12:41 PM · Restricted Project
asaadaldien updated the diff for D105244: Add linalg.mmt4d named op.
rebase
Jul 1 2021, 10:58 AM · Restricted Project
asaadaldien added inline comments to D105063: [mlir][tosa] Add quantized lowering for matmul and fully_connected.
Jul 1 2021, 9:59 AM · Restricted Project
Jun 30 2021
Jun 30 2021
asaadaldien updated the summary of D105244: Add linalg.mmt4d named op.
Jun 30 2021, 4:07 PM · Restricted Project
asaadaldien updated the summary of D105244: Add linalg.mmt4d named op.
Jun 30 2021, 4:07 PM · Restricted Project
asaadaldien requested review of D105244: Add linalg.mmt4d named op.
Jun 30 2021, 4:06 PM · Restricted Project
Add linalg.batch_matvec named op
Jun 30 2021, 11:37 AM · Restricted Project
asaadaldien updated the diff for D104739: Add linalg.batch_matvec named op.
Rebase...
Jun 30 2021, 11:17 AM · Restricted Project
asaadaldien updated the diff for D104739: Add linalg.batch_matvec named op.
Generate YAML changes by running mlir.dialects.linalg.opdsl.dump
Jun 30 2021, 9:40 AM · Restricted Project
Jun 27 2021
Jun 27 2021
asaadaldien added inline comments to D104884: Order parallel indices before transposing the input in multireductions.
Jun 27 2021, 11:19 PM · Restricted Project
asaadaldien updated the diff for D104739: Add linalg.batch_matvec named op.
Rebase...
Jun 27 2021, 9:59 PM · Restricted Project
Jun 22 2021
Jun 22 2021
asaadaldien retitled D104739: Add linalg.batch_matvec named op from Add linalg.batch_matvec named opSimilarly to batch_mat vec outer most dim is a batching dimand this op does |b| matrix-vector-products : C[b, i] = sum_k(A[b, i, k] * B[b, k]) to Add linalg.batch_matvec named op.
Jun 22 2021, 12:56 PM · Restricted Project
asaadaldien retitled D104739: Add linalg.batch_matvec named op from Add linalg.batch_matvec named op
Similarly to batch_mat vec outer most dim is a batching dim
and this op does |b| matrix-vector-products :
C[b, i] = sum_k(A[b, i, k] * B[b, k]) to Add linalg.batch_matvec named opSimilarly to batch_mat vec outer most dim is a batching dimand this op does |b| matrix-vector-products : C[b, i] = sum_k(A[b, i, k] * B[b, k]).
Jun 22 2021, 12:56 PM · Restricted Project
asaadaldien requested review of D104739: Add linalg.batch_matvec named op.
Jun 22 2021, 12:55 PM · Restricted Project
Jun 21 2021
Jun 21 2021
asaadaldien committed rG7e2d672a672c: Add polynomial approximation for trigonometric sine and cosine functions (authored by asaadaldien).
Add polynomial approximation for trigonometric sine and cosine functions
asaadaldien closed D104582: Add polynomial approximation for trigonometric sine and cosine functions.
Jun 21 2021, 1:00 PM · Restricted Project
asaadaldien updated the diff for D104582: Add polynomial approximation for trigonometric sine and cosine functions.
assert cosOp&sinOp types and fix clang-tidy warning
Jun 21 2021, 9:51 AM · Restricted Project
asaadaldien updated the summary of D104582: Add polynomial approximation for trigonometric sine and cosine functions.
Jun 21 2021, 9:42 AM · Restricted Project
asaadaldien updated the diff for D104582: Add polynomial approximation for trigonometric sine and cosine functions.
consistant constant / utility func naming
Jun 21 2021, 9:34 AM · Restricted Project
Jun 18 2021
Jun 18 2021
asaadaldien added a reviewer for D104582: Add polynomial approximation for trigonometric sine and cosine functions: ezhulenev.
Jun 18 2021, 4:55 PM · Restricted Project
asaadaldien requested review of D104582: Add polynomial approximation for trigonometric sine and cosine functions.
Jun 18 2021, 4:54 PM · Restricted Project
Jun 10 2021
Jun 10 2021
asaadaldien committed rG20daedacca80: 2d Arm Neon sdot op, and lowering to the intrinsic. (authored by Benoit).
2d Arm Neon sdot op, and lowering to the intrinsic.
Jun 10 2021, 2:36 PM · Restricted Project
Jun 9 2021
Jun 9 2021
asaadaldien committed rGb9d7ffd9cf5f: Folds linalg.pad_tensor with zero padding (authored by asaadaldien).
Folds linalg.pad_tensor with zero padding
Jun 9 2021, 3:43 PM · Restricted Project
asaadaldien requested review of D103984: Folds linalg.pad_tensor with zero padding.
Jun 9 2021, 12:06 PM · Restricted Project
Jun 4 2021
Jun 4 2021
asaadaldien added a comment to D103724: Revert "Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns".
Apologies for pushing the revert without justifications. The original commit wasn't ready to land at the time it landed.
Jun 4 2021, 4:42 PM · Restricted Project
asaadaldien committed rGbba8d8c1868d: Revert "Add memref.dim canonicalization patterns to… (authored by asaadaldien).
Revert "Add memref.dim canonicalization patterns to…
asaadaldien added a reverting change for D103715: Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns: rGbba8d8c1868d: Revert "Add memref.dim canonicalization patterns to….
Jun 4 2021, 3:44 PM · Restricted Project
asaadaldien closed D103724: Revert "Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns".
Jun 4 2021, 3:44 PM · Restricted Project
asaadaldien requested review of D103724: Revert "Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns".
Jun 4 2021, 3:41 PM · Restricted Project
asaadaldien added a reverting change for D103715: Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns: D103724: Revert "Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns".
Jun 4 2021, 3:41 PM · Restricted Project
asaadaldien committed rGa52959401d95: Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns (authored by asaadaldien).
Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns
asaadaldien closed D103715: Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns.
Jun 4 2021, 1:41 PM · Restricted Project
asaadaldien requested review of D103715: Add memref.dim canonicalization patterns to TilingCanonicalizationPatterns.
Jun 4 2021, 1:33 PM · Restricted Project
May 12 2021
May 12 2021
asaadaldien accepted D102301: [mlir][Linalg] Add interface methods to get lhs and rhs of contraction op..
May 12 2021, 11:07 AM · Restricted Project
May 3 2021
May 3 2021
asaadaldien accepted D101795: [mlir] Linalg: add vector transfer lowering patterns to the contraction lowering.
May 3 2021, 4:09 PM · Restricted Project
Apr 30 2021
Apr 30 2021
asaadaldien committed rG499e89fc9119: Add patterns to lower vector.multi_reduction into a sequence of vector.reduction (authored by asaadaldien).
Add patterns to lower vector.multi_reduction into a sequence of vector.reduction